2026-09-19 AI 日报
统计日时区：America/Los_Angeles · 已结算
cactus-compute/needle · trycua/cua · higgsfield-ai/higgsfield
开源项目
- cactus-compute/needle
Automation foundation model for tiny devices: 2-bit, 8-29 MB, tool calls, structured extraction and embeddings on phones, wearables, smart homes, robots, cars…
- trycua/cua
Scale computer-use 2.0 with open-source drivers, cross-OS fleets, and benchmarks for training, evaluation, and data generation.
- higgsfield-ai/higgsfield
Fault-tolerant, highly scalable GPU orchestration, and a machine learning framework designed for training models with billions to trillions of parameters
- krillinai/OpenCreator
Formerly KrillinAI. Open-source AI workspace for creators, powered by Codex. Create videos, images, voice, avatars, translations, and edits with Agents in one…
- alibaba/open-code-review
Secure, fast, efficient, battle-tested at Alibaba's scale. Hybrid architecture code review tool: deterministic pipelines + LLM Agent, precise line-level…
- stablyai/orca
Orca is the ADE for working with a fleet of parallel agents. Run any coding agent with your own subscription. Available on desktop, mobile and remote runtime.
- yynxxxxx/Codex-X
OpenAI Codex 桌面端/CLI 的可视化管理工具，具有Provider/API 切换、会话同步、提示词注入、Skills/MCP 管理、TOML 配置可视化的跨平台工具。
- addyosmani/agent-skills
Production-grade engineering skills for AI coding agents.
- multica-ai/multica
Make humans and AI agents work as one team — open-source and self-hostable.
- Tencent/WeKnora
Open-source LLM knowledge platform: turn raw documents into a queryable RAG, an autonomous reasoning agent, and a self-maintaining Wiki.
论文
本统计日暂无条目。
行业新闻
- I built non-autoregressive decision models with RL a year ago
- Lawsuit says Anthropic, OpenAI and others made illegal agreement on AI slowdown
- The Hugging Face Hack Wasn't What It Was Cracked Up to Be
- OpenAI and Anthropic oversold AI security breaches
- Orchestrating Claude Code Agents: The Chief of Staff Pattern
- Partnering with Accenture on Embedded Evaluation
- Lawsuit: Illegal Agreement of Anthropic, OpenAI, SpaceXAI, Google on AI Slowdown
- The Case of Elias Thorne, Imaginary Man AI Chatbots Are Obsessed With
社区动态
- Meet Qwen3.8-LiveTranslate, Qwen's next-generation real-time simultaneous interpretation model!
Meet Qwen3.8-LiveTranslate, Qwen's next-generation real-time simultaneous interpretation model! 📢 Built on an Interleave architecture, it improves…
- What are you vibe coding this weekend?
What are you vibe coding this weekend?
- A great example of practical AI solving real-world problems!
A great example of practical AI solving real-world problems! Faster than ever. Thanks for building with Qwen. @cerebras Try it out for yourself! 🏠
- With Gemini 4, bench goes up.
They claimed open-weight models are dangerous but the benchmarks say otherwise. Source
- I wanted to watch a neural network learn [P]
I wanted to really see how a neural network learns different functions, so built an interactive demo. You can change the architecture of the network and the…
- 🚨 Gemini 4 Pro vs GPT-6 Astra Pro (Xbox Controller SVG) I added some more examples too, Gemini 4 Pro and Astra look…
🚨 Gemini 4 Pro vs GPT-6 Astra Pro (Xbox Controller SVG) I added some more examples too, Gemini 4 Pro and Astra look near perfect but the logo is slightly off…
- Please stop with the FP4 inference engines for the love of god
Every day there's a new post of some optimized config or new inference engine that is just super good at one specific thing and their claims make sense. And…
- General warning about Clore.AI
Hello, I know some of us may be tempted to rent out our expensive GPUs to recoup some of the cost of self-hosting, and it should be obvious that this can be a…
- Found this in a Chinese Grade 3 IT textbook,it teaches kids how to ask DeepSeek questions
Image 1 is the original textbook page I happened to come across, and Image 2 is the English version I asked GPT to make. AI has already become part of many…
- A SLM Benchmark
Empirical evaluation frameworks and reproducible datasets for Small Language Model (SLM) Task Governance and Deterministic Orchestration.…
官方发布
- Introducing Grok Voice Transcribe 2.0
- Qwen3.8-LiveTranslate: Names the speaker. Carries the meaning.
Simultaneous interpretation is not only about translating fast — it must also hear clearly and translate accurately. Qwen3.8-LiveTranslate rebuilds real-time…
- Qwen/Qwen-Image-2.1
text-to-image
- Introducing Gemini 3.8 Live and 3.8 Live Extended Thinking
Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking are our most advanced live dialogue models yet, built for natural conversation.
- Partnering with Accenture on embedded evaluation
- Qwen3.8-Omni-Flash: Omni Senses. Agentic Delivery.
Today, we are launching Qwen3.8-Omni-Flash, our next-generation native omnimodal model. Its core objective is to strengthen agent capabilities in real-world…
- The Compliance API local session endpoints now also return transcripts of Claude in Chrome sessions (product_surface…
The Compliance API local session endpoints now also return transcripts of Claude in Chrome sessions (product_surface value claude_in_chrome), in beta for…
- For cache diagnostics, a response to a request that sends the cache-diagnosis-2026-04-07 beta header now always…
For cache diagnostics, a response to a request that sends the cache-diagnosis-2026-04-07 beta header now always includes the diagnostics field. The field is…
- Why client SDK generation belongs in the open
Google has partnered with Speakeasy to open-source their OpenAPI code generation suite under the AGPLv3 license, a strategic move prompted by the sudden…
- ChatGPT for Word
ChatGPT is now available in Microsoft Word. You can draft from notes, summarize a document, revise selected text, and adjust headings and formatting from the…